Q&A Series: Harnessing the power of AI in media buying

By Ran Milo

Interviewing one of our in house media buying specialists, we take a deeper dive into how AI helps amplify performance and results.

About Matt: Matt is a proud Media Buying Specialist and Performance Manager at Bidalgo. He works with leading mobile advertisers and uses Bidalgo’s AI powered platform for strong cross channel performance.

Off the bat, how does AI make you a better media buyer?

In a nutshell, AI amplifies my results. It looks at my performance on a minute by minute, hour by hour basis and makes it even stronger. It takes my ongoing campaign activity and amplifies it in unique ways- ways that I might not have been able to achieve on my own.

AI amplifies my work by: noticing trends, bringing forward potential insights I may have overlooked, automating my menial day to day tasks so I can focus on testing new assets and audiences. Most importantly, it provides me with creative insights so I can improve my campaign creatives – I find this critical in today’s performance marketing landscape. Bidalgo’s AI is active across multiple channels and  allows me to focus on specific audiences I want to reach, rather than on how to succeed on a specific channel.

What are the main fears around using AI in media buying?

There is a strange misconception that AI takes control away from manual activity. In reality AI should be used as a building block for said activity. From my experience, AI is the constant stable portion of activity, spends consistent amounts, and brings in steady results, freeing up time for performance managers and media buyers to test new activity and to improve their own methods. AI can then take newly acquired data and run with the most successful portions of it for optimized results. There is also another strange thought out there that AI will take over media buying leaving nothing for us human media buyers to do! My experience was quite the opposite-  AI has allowed me to increase the scale of my managed campaigns, it has surfaced new insights I can use to improve my performance and has allowed me to focus my time on the meaningful work. For me, AI is like another team member I’m managing. I can give it tasks and learn from it, yet I need to give it the right directions and feed it with the right assets. AI brings amplified performance to an already strong team. This is one of my favorite things about it.

So what actually is your favorite thing about using AI in your campaigns?

One of the hardest things to do as a scaling company is to manage continuous amounts of revenue whilst finding a way to accumulate insights and learn from massive amounts of data. AI achieves this in one swoop – it acts as a team member, communicates relevant intel and reveals data you could not have discovered manually. Like any strong team member, it will provide strong consistency and commitment.

Matt Skurnick


With the right AI in place, performance managers get to see startling trends as they delve into its data.  For example they can better understand why a certain conversion spec was chosen, why a certain lookalike works with a specific creative, why spend elevates or drops by X% on a specific day and/or time. As performance managers, we can collect this data and get new ideas to further test and expand.  For example, a recent AI test assessed CPM on an hour by hour basis and told me the best time for spend- it allowed me to detect a performance pattern showcasing stronger weekend performance for a specific app- which would have been significantly harder to identify had I used manual activity alone. Based on this intel I was able to focus more effort on weekend performance and bring the client improved results.

What can AI do for KPIs?

On a pretty consistent basis, AI has been able to hit the same or higher ROAS or Lower CPAs depending on the KPI we are optimizing towards.  This is due to its consistent testing – analyzing data on a minute by minute basis compared to manual attempts which would be day by day. At Bidalgo we review AI prior to spending, we let it do its thing- namely hitting our primary geos with consistent spend, so we as performance managers can retain some kind of flexibility and explore new opportunities.

What’s your advice for those getting started with AI-powered media buying?

Just like any activity, you need to start small. AI will not have the historical track record to work on from the get-go (although you can easily sync your historical data to it to get an immediate start). You need to start on a certain level and work your way up. Making drastic changes early on can harm activity. Remember that AI is a team member – you need to check up on it, add assets on a regular basis and give it some TLC. Once you treat it right, it will do what it does best-  deliver results. It’s a plant that needs to grow.

Tell us about how your performance has changed since using Bidalgo AI.

Within months of using Bidalgo’s Creative AI I have seen ROAS D7 increase by up to 40% and ARPPU (Average Revenue Per Paying User) by 60% thanks to AI alone. Using AI has allowed me to eliminate guesswork and iterations involved with developing ad creatives and focus a lot more on results.

How do you see this kind of AI changing performance marketing in the next 10 years?

Two words- The Jetsons.  Actually I think that in the future, AI will be able to test new features that are released by marketing channels without the need for past data.  I also think that Bidalgo’s Creative AI will be able to make creative iterations (changing color, movement, facial expressions, characters, etc) on marketing material and run small scale creative testing on it’s own without any human intervention.

Published on November 27, 2018
Written by
Ran Milo

VP Marketing @ Bidalgo.

Get Started With Bidalgo Today